Olivier Lartillot
Department of Architecture, Design and Media Technology, Aalborg University
Date: Wednesday 11 May 2016
Time: 12.45 – 14.00
Place: Rendsburggade 14, Room 4.315
Abstract
The aim of this research is to extract all repetitions of motifs in music, through a search for repeated substrings. One particular challenge is to describe the pattern repetition exhaustively but in a most compact way. We introduce a new method for online closed pattern mining and address the important issue, quite common in music, where a pattern repeats many times successively. This method is generalised to the search for heterogeneous patterns in multidimensional strings, with musical transformations, and in polyphony where multiple notes appear at the same time. More generally, the analysis of music is modelled as a network of interdependent modules.
Bio
Olivier Lartillot is member of the Music Informatics and Cognition Group. He is a researcher in computational music analysis, in both domains of audio and score analysis. As an Academy of Finland research fellow, he designed Matlab software for music analysis (MIRtoolbox) as well as for music therapy improvisation analysis. He participates to the European project Learning to Create.